Power transmission tower geographical environment remote sensing image analysis method, system, device and medium

Abstract

This invention discloses a method, system, equipment, and medium for remote sensing image analysis of the geographical environment of power transmission towers, belonging to the field of remote sensing image analysis technology. The method includes: obtaining power transmission tower detection results based on remote sensing images; if a rectangular frame of the tower body exists in the detection results, obtaining geographical environment segmentation results based on the remote sensing images; if a geographical environment mask exists in the segmentation results, performing mask region matching and mask region expansion based on the power transmission tower detection results and the geographical environment segmentation results to obtain mask regions of different levels; quantitatively analyzing the different levels of mask regions based on the degree of influence of environmental types to obtain different levels of geographical environment influence factors; and determining the degree of influence of the geographical environment on the power transmission towers based on the different levels of geographical environment influence factors. This achieves quantitative analysis of the influence of the surrounding geographical environment on power transmission towers, greatly improving the efficiency of power line inspection work.

Description

Technical Field

[0001] This invention relates to the field of remote sensing image analysis technology, and in particular to a method, system, equipment, and medium for analyzing the geographical environment of power transmission towers using remote sensing images. Background Technology

[0002] Currently, remote sensing image analysis methods for power transmission towers or geographical environments often employ traditional machine vision algorithms such as SVM and Bayes, or deep learning algorithms such as YOLO and Faster-RCNN. However, due to the large resolution of satellite remote sensing images, the characteristics of power transmission towers are not obvious, and the surrounding geographical environment is complex and variable. Therefore, it is difficult to analyze the degree of influence of the geographical environment on power transmission towers by directly using traditional machine vision algorithms or deep learning algorithms. Summary of the Invention

[0003] In view of this, the purpose of this invention is to overcome the shortcomings of the prior art and provide a method, system, device and medium for analyzing remote sensing images of the geographical environment of power transmission towers. This method is used to assess the degree of influence of the geographical environment on power transmission towers by performing remote sensing detection of power transmission towers, geographical environment mask segmentation, mask area matching and expansion, geographical environment quantitative analysis and determination of influence level of remote sensing images.

[0004] This invention provides the following technical solution:

[0005] In a first aspect, the present invention proposes a method for analyzing remote sensing images of the geographical environment of power transmission towers, including:

[0006] Remote sensing imagery is acquired and input into a remote sensing detection algorithm for power transmission towers to obtain tower detection results. If a tower body rectangle exists in the tower detection results, the remote sensing imagery is input into a geographic environment remote sensing segmentation algorithm to obtain geographic environment segmentation results. If a geographic environment mask region exists in the geographic environment segmentation results, mask region matching and mask region expansion are performed based on the tower detection results and the geographic environment segmentation results to obtain the first-level mask region, second-level mask region, and third-level mask region corresponding to each tower body rectangle. The first-level masking area, second-level masking area, and third-level masking area corresponding to each of the tower body rectangles are quantitatively analyzed based on the degree of influence of environmental type, to obtain the first-level geographical environment influence factor, second-level geographical environment influence factor, and third-level geographical environment influence factor corresponding to each of the tower body rectangles; the degree of influence of geographical environment on the transmission tower corresponding to each of the tower body rectangles is determined based on the first-level geographical environment influence factor, second-level geographical environment influence factor, and third-level geographical environment influence factor corresponding to each of the tower body rectangles.

[0007] Secondly, this invention proposes a remote sensing image analysis system for the geographical environment of power transmission towers, comprising:

[0008] The target detection module is used to acquire remote sensing images and input the remote sensing images into the transmission tower remote sensing detection algorithm to obtain the transmission tower detection results.

[0009] The mask segmentation module is used to input the remote sensing image into the geographic environment remote sensing segmentation algorithm to obtain the geographic environment segmentation result if the tower body rectangle exists in the tower detection result.

[0010] The matching expansion module is used to match and expand the mask region according to the transmission tower detection result and the geographical environment segmentation result if there is a geographical environment mask region in the geographical environment segmentation result, so as to obtain the first-level mask region, the second-level mask region and the third-level mask region corresponding to the rectangular frame of each tower body;

[0011] The quantitative analysis module is used to perform quantitative analysis on the first-level mask area, second-level mask area and third-level mask area corresponding to the rectangular frame of each tower body based on the degree of influence of environmental type, so as to obtain the first-level geographical environment influence factor, second-level geographical environment influence factor and third-level geographical environment influence factor corresponding to the rectangular frame of each tower body.

[0012] The degree determination module is used to determine the degree of influence of the geographical environment on the transmission towers corresponding to the rectangular frames of each tower body, based on the first-level geographical environment influence factor, the second-level geographical environment influence factor, and the third-level geographical environment influence factor corresponding to the rectangular frames of each tower body.

[0013] Thirdly, the present invention proposes a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the remote sensing image analysis method for the geographical environment of transmission towers described in the first aspect.

[0014] Fourthly, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the remote sensing image analysis method for the geographical environment of transmission towers described in the first aspect.

[0015] The remote sensing image analysis method for the geographical environment of power transmission towers disclosed in this invention constructs a fully automated remote sensing image analysis process, including remote sensing detection of power transmission towers, geographical environment mask segmentation, mask region matching and expansion, quantitative analysis of the geographical environment, and determination of the degree of influence. It overcomes the problems of low detection accuracy and difficulty in quantifying environmental impact caused by small targets and complex backgrounds in high-resolution remote sensing images, which are caused by traditional machine vision algorithms and general deep learning models. It realizes the quantitative analysis of the degree of influence of the surrounding geographical environment on power transmission towers, and greatly improves the efficiency of power inspection work. Attached Figure Description

[0016] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope of protection of the present invention. In the various drawings, similar components are numbered similarly.

[0017] Figure 1 This embodiment shows a flowchart of the remote sensing image analysis method for the geographical environment of power transmission towers.

[0018] Figure 2 A schematic diagram of the network structure of the remote sensing detection algorithm for transmission towers proposed in this embodiment is shown.

[0019] Figure 3 A schematic diagram of the mask region distribution proposed in this embodiment is shown;

[0020] Figure 4 This embodiment illustrates a flowchart of the quantitative analysis of different levels of mask regions proposed in this example;

[0021] Figure 5 A schematic diagram of the structure of the remote sensing image analysis system for the geographical environment of power transmission towers proposed in this embodiment is shown.

[0022] Explanation of reference numerals in the attached diagram:

[0023] 500 - Remote sensing image analysis system for the geographical environment of power transmission towers; 501 - Target detection module; 502 - Mask segmentation module; 503 - Matching expansion module; 504 - Quantitative analysis module; 505 - Degree determination module. Detailed Implementation

[0024] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0025] The components of the embodiments of the invention described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0026] In the following, the terms “comprising,” “having,” and their cognates, which may be used in various embodiments of the invention, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as excluding, firstly, the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more features, numbers, steps, operations, elements, components, or combinations thereof.

[0027] Furthermore, the terms "first," "second," and "third" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0028] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of the invention pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be interpreted as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of the invention.

[0029] Example 1

[0030] This disclosure provides a method for analyzing remote sensing images of the geographical environment of power transmission towers. The method involves remote sensing detection of power transmission towers, geographical environment masking segmentation, masking region matching and expansion, quantitative analysis of the geographical environment, and determination of the degree of influence of the geographical environment on power transmission towers.

[0031] Please see Figure 1 The method for analyzing the geographical environment of power transmission towers using remote sensing images includes steps S101 to S105, and each step is described in detail below.

[0032] Step S101: Acquire remote sensing images and input the remote sensing images into the power transmission tower remote sensing detection algorithm to obtain the power transmission tower detection results.

[0033] In this embodiment, high-resolution remote sensing images are acquired and input into a power transmission tower remote sensing detection algorithm for target detection, resulting in power transmission tower detection results. If a power transmission tower exists in the remote sensing image, the power transmission tower detection results will show at least one tower body rectangle and its corresponding tower base rectangle.

[0034] In one specific embodiment, the remote sensing detection algorithm for power transmission towers is an improved YOLO11 detection algorithm. The improved YOLO11 detection algorithm includes a backbone network, a feature pyramid network, and a head detection network. The feature pyramid network includes a first self-attention convolutional module, a second self-attention convolutional module, a first convolutional layer, a second convolutional layer, a first SwinTransformer module, a second SwinTransformer module, and a third SwinTransformer module. The head detection network includes a first classification regression module, a second classification regression module, a third classification regression module, a fourth classification regression module, a fifth classification regression module, and a sixth classification regression module. The system comprises a regression module, a first interest alignment module, a second interest alignment module, and a third interest alignment module. Step S101 includes: extracting features from the remote sensing image through the backbone network to obtain a first layer feature map, a second layer feature map, and a third layer feature map; inputting the third layer feature map and the second layer feature map into the first self-attention convolution module to obtain a first self-attention convolution feature map, and concatenating the first self-attention convolution feature map with the second layer feature map to obtain a first fused map; inputting the first fused map and the first layer feature map into the second self-attention convolution module to obtain a second self-attention convolution feature map, and concatenating the second self-attention convolution feature map with the first layer feature map. The first layer feature maps are concatenated to obtain a second fused map; the second fused map is input into the first convolutional layer to obtain a first feature map, and the first feature map is concatenated with the first fused map to obtain a third fused map; the third fused map is input into the second convolutional layer to obtain a second feature map, and the second feature map is concatenated with the third layer feature map to obtain a fourth fused map; the second fused map is input into the first SwinTransformer module to obtain a first fine-grained feature map, the third fused map is input into the second SwinTransformer module to obtain a second fine-grained feature map, and the fourth fused map is input into the third SwinTransformer module. The `inTransformer` module obtains a third fine-grained feature map. The first, second, and third fine-grained feature maps are then input into the first, second, and third classification / regression modules, respectively, to obtain first, second, and third tower body detection results. These results are then input into the first, second, and third interest alignment modules, respectively, to obtain a first alignment feature map, a second alignment feature map, and a third alignment feature map.The first alignment feature map, the second alignment feature map, and the third alignment feature map are respectively input into the fourth classification regression module, the fifth classification regression module, and the sixth classification regression module to obtain the detection results of the first tower base, the second tower base, and the third tower base. Based on the detection results of the first tower body, the second tower body, the third tower body, the first tower base, the second tower base, and the third tower base, the detection result of the transmission tower is obtained.

[0035] In this embodiment, a remote sensing detection algorithm for transmission towers based on improved YOLO11 detection is used to detect targets in remote sensing images.

[0036] Among them, a self-attention mechanism is introduced, which enables high-level semantic features to focus on low-level detailed regions (such as pole edges and tower base outlines), enhances the model's attention to key parts, and solves the problem of insufficient feature fusion.

[0037] SwinTransformer replaces the ordinary feature aggregation module and uses a window attention mechanism to capture long-distance dependencies, effectively modeling the overall structural morphology of towers. It is especially suitable for the characteristics of targets with variable orientations and complex poses in remote sensing images.

[0038] The ROI Align module performs pairwise detection based on the output of the Classification Regression module, ensuring spatial consistency between the tower body and its base, preventing misjudgments such as towers without bases or misaligned matches, and improving the physical plausibility of the detection results. Simultaneously, it enhances the high-level semantic feature representation of the target and effectively controls the computational efficiency of target features.

[0039] For example, please see Figure 2 The backbone network includes Conv3, Conv4, and Conv5. Remote sensing images are input into the backbone network to obtain the first-layer feature map F3 corresponding to the output of Conv3, the second-layer feature map F4 corresponding to the output of Conv4, and the third-layer feature map F5 corresponding to the output of Conv5.

[0040] Furthermore, the third-layer feature map F5 and the second-layer feature map F4 are input into the first self-attention convolution module (SACM) to obtain the first self-attention convolution feature map SACF4, and the first self-attention convolution feature map SACF4 is concatenated with the second-layer feature map F4 to obtain the first fused map ~P4.

[0041] Furthermore, the first fused image ~P4 and the first layer feature map F3 are input into the second self-attention convolution module to obtain the second self-attention convolution feature map SACF3, and the second self-attention convolution feature map SACF3 is concatenated with the first layer feature map F3 to obtain the second fused image P3.

[0042] Furthermore, the second fused image P3 is input into the first convolutional layer Conv to obtain the first feature map PAF4 with the same resolution as ~P4. The first feature map PAF4 is then concatenated with the first fused image ~P4 to obtain the third fused image P4.

[0043] Furthermore, the third fusion map P4 is input into the second convolutional layer to obtain the second feature map PAF5 with the same resolution as F5. The second feature map PAF5 is then concatenated with the third layer feature map F5 to obtain the fourth fusion map P5.

[0044] Furthermore, the second fused image P3, the third fused image P4, and the fourth fused image P5 are subjected to high-level fine-grained feature extraction through the first SwinTransformer module, the second SwinTransformer module, and the third SwinTransformer module, respectively, to obtain cross-level features from top to bottom and bottom to top, including the first fine-grained feature map T3, the second fine-grained feature map T4, and the third fine-grained feature map T5.

[0045] Furthermore, the first fine-grained feature map T3, the second fine-grained feature map T4, and the third fine-grained feature map T5 are input into the first classification regression module, the second classification regression module, and the third classification regression module, respectively, to obtain the first tower body detection result, the second tower body detection result, and the third tower body detection result.

[0046] Furthermore, the detection results of the first tower body, the second tower body, and the third tower body are input into the first interest alignment module, the second interest alignment module, and the third interest alignment module, respectively, to obtain the first alignment feature map, the second alignment feature map, and the third alignment feature map, which are ROI Align feature maps of uniform size.

[0047] Furthermore, the first alignment feature map, the second alignment feature map, and the third alignment feature map are input into the fourth classification regression module, the fifth classification regression module, and the sixth classification regression module, respectively, to obtain the detection results of the first tower base, the second tower base, and the third tower base, thus completing the paired detection task of the tower body and the tower base.

[0048] Furthermore, the inspection results of the first tower body, the second tower body, the third tower body, the first tower base, the second tower base, and the third tower base are integrated to obtain the transmission tower inspection results.

[0049] It should be noted that the training process for the remote sensing detection model of transmission towers is not limited in this embodiment.

[0050] In addition, the remote sensing images input to the transmission tower remote sensing detection algorithm can be cropped into sliding windows according to a preset resolution size (e.g., 1000×1000), and the sliding window sub-images are sequentially input into the transmission tower remote sensing detection algorithm for tower detection.

[0051] Specifically, the coordinates of multiple tower body rectangles and corresponding multiple tower base rectangles in each sliding window image set are mapped back to the coordinates of the high-resolution remote sensing image, resulting in the mapped multiple tower body rectangles and corresponding multiple tower base rectangles.

[0052] In high-resolution remote sensing images, power transmission towers occupy a very small proportion (often only tens of pixels), and the entire image input is easily obscured by the background. By using a sliding window to decompose a large image into local sub-images, the proportion of the target in the local field of view is increased, which is beneficial for the network to extract features and significantly improves the detection rate of small targets.

[0053] For the multiple mapped tower body rectangles, according to the confidence filtering principle, tower body rectangles with a confidence level below the threshold (which can be 0.35) are deleted; according to the area non-maximum suppression (ANMS) algorithm, tower body rectangles with a variation overlap rate above the threshold (which can be 0.85) are deleted, resulting in multiple tower body rectangles; correspondingly, due to paired target detection, the tower base rectangles corresponding to each tower body rectangle can be obtained.

[0054] For Area Non-Maximum Suppression (ANMS), firstly, all rectangles are sorted from largest to smallest area. Then, the rectangle with the largest area is selected as the registration rectangle. The variation overlap rate (VAR) of the remaining rectangles is calculated with the registration rectangle, so each remaining rectangle corresponds to a VAR. If the VAR exceeds a threshold (which can be 0.85), the corresponding remaining rectangle is deleted; otherwise, it is retained. Next, the rectangle with the largest area is selected again from the remaining rectangles as the registration rectangle. The VAR of the other remaining rectangles is calculated with the registration rectangle. Remaining rectangles with VARs exceeding the threshold are deleted, while those with VARs below the threshold are retained. This process is repeated until none of the selected registration rectangles have a VAR exceeding the threshold. Finally, all selected registration rectangles are used as the output rectangles of the final algorithm.

[0055] Regarding the variation overlap rate, assume the top left point of rectangle ABCD bottom right Top left point of rectangle EFGH bottom right Then the variation overlap rate between rectangle ABCD and rectangle EFGH can be expressed as: , , , ,in, Indicates the overlap rate of variations. This represents the area of ​​the intersection of two rectangles.

[0056] Step S102: If the tower body rectangle exists in the tower detection result, the remote sensing image is input into the geographic environment remote sensing segmentation algorithm to obtain the geographic environment segmentation result.

[0057] In this embodiment, if the tower body rectangle exists in the tower detection result, the remote sensing image is input into the geographic environment remote sensing segmentation algorithm to obtain the geographic environment segmentation result; if the tower body rectangle does not exist in the tower detection result, it is considered that there is no tower in the high-resolution remote sensing image, and the tower remote sensing detection of the next high-resolution remote sensing image needs to be started.

[0058] The geographic environment segmentation results generally include multiple tower base mask areas, multiple geographic environment mask areas and corresponding geographic environment types. The geographic environment types specifically include buildings, vegetation, water bodies, floating objects, etc.

[0059] It should be noted that the geographic environment segmentation model is obtained by training the SFNET semantic segmentation algorithm. The entire SFNET network architecture includes a bottom-up pathway as an encoder and a top-down pathway as a decoder. The encoder has the same backbone as the image classification, and the final fully connected layer is replaced by a context modeling module (PPM). The decoder is a typical FPN structure equipped with FAM.

[0060] Backbone: A ResNet series pre-trained on ImageNet is used as the backbone network for main feature extraction. The last fully connected layer for classification is removed. In addition, all backbones have four stages, and the first convolutional layer of each stage is a convolution with a stride of 2 to downsample the feature maps to improve computational efficiency and obtain a larger receptive field.

[0061] Context Module: The PPM module is used to capture long-range context information. Since the output of PPM has the same resolution feature map as the last residual module, PPM and the last residual module are used as the final stage of FPN.

[0062] The Aligned FPN Decoder: This decoder extracts feature maps from the encoder and uses the aligned feature pyramid for final scene parsing. In the decoder, the original bilinear upsampling operation is replaced with a FAM module to align the top-level feature maps, and then the pyramid is added to the corresponding bottom-level feature maps using an Add operation. Finally, the feature maps of each stage in the decoder are upsampled to 1 / 4 of the original input image, and these feature maps are concatenated for the final prediction. To address the alignment issues between different levels, a FAM module is also used to replace the original upsampling operation.

[0063] Understandably, by conditionally triggering the geographic environment segmentation process (which is only executed when a tower is detected), unnecessary computational resource consumption can be avoided, overall analysis efficiency can be improved, and it is suitable for large-scale batch processing of remote sensing images.

[0064] As an example, for the geographic environment segmentation process, a high-resolution remote sensing image is cropped into a sliding window at a preset resolution size (e.g., 600×600) to obtain a set of second sliding window sub-images. Weights are assigned to each second sliding window sub-image according to a two-dimensional normal probability distribution, resulting in sliding window weighted sub-images with larger weights in the central region and smaller weights in the edge region. The second sliding window images are then input into a geographic environment remote sensing segmentation model trained using the SFNET semantic segmentation algorithm to obtain geographic environment segmentation mask sub-images for the second sliding window images. The coordinates of all sliding window weighted sub-images and all geographic environment segmentation mask sub-images are mapped back to the coordinates of the high-resolution remote sensing image, resulting in multiple mapped sliding window weighted sub-images and multiple mapped geographic environment segmentation mask sub-images.

[0065] The geographic environment segmentation mask image is obtained by normalizing multiple mapped sliding window weighted sub-images and multiple mapped geographic environment segmentation mask sub-images using the mask normalization formula.

[0066] The geographic environment segmentation mask image represents the geographic environment segmentation result, containing multiple geographic environment mask regions and corresponding geographic environment types, including tower bases, buildings, vegetation, water bodies, floating objects, etc. The mask normalization formula is as follows: In the formula, Representing coordinates in high-resolution remote sensing images The mask value of the geographic environment segmentation mask image corresponding to the location; Representing coordinates in high-resolution remote sensing images The location corresponding to the first The mask value of a mapped geographic environment segmentation mask sub-image. If it does not exist, the default value is 0; Representing coordinates in high-resolution remote sensing images The location corresponding to the first The weight values ​​of the mapped sliding window weighted sub-images If it does not exist, the default value is 0; This indicates the number of images in the second sliding window.

[0067] Step S103: If there is a geographic environment mask region in the geographic environment segmentation result, then mask region matching and mask region expansion are performed according to the transmission tower detection result and the geographic environment segmentation result to obtain the first-level mask region, second-level mask region and third-level mask region corresponding to the rectangular frame of each tower body.

[0068] In this embodiment, if there is a geographic environment mask region in the geographic environment segmentation result, the spatial location of the transmission tower and the spatial distribution of the geographic environment region and the tower base region are analyzed based on the transmission tower detection result and the geographic environment segmentation result. This is to achieve the matching of each tower body rectangle in the transmission tower detection result with its corresponding tower base mask region. Furthermore, the tower base mask region is expanded in multiple levels to obtain the first-level mask region, the second-level mask region and the third-level mask region corresponding to each tower body rectangle. Thus, a multi-scale impact assessment framework is constructed using different levels of mask regions.

[0069] Meanwhile, if there is no geographic environment mask area in the geographic environment segmentation result, it is considered that there are no geographic environment influence factors of power transmission towers in the high-resolution remote sensing image, and the remote sensing detection of power transmission towers in the next high-resolution remote sensing image is started.

[0070] In one specific embodiment, step S103 includes: matching the detection results of the transmission tower and the geographical environment segmentation results to obtain the tower base mask area corresponding to each of the tower body rectangles; for each of the tower body rectangles, using the tower base mask area corresponding to the tower body rectangle as the first-level mask area corresponding to the tower body rectangle; expanding the tower base mask area outward according to a first preset ratio to obtain the second-level mask area corresponding to the tower body rectangle; and expanding the tower base mask area outward according to a second preset ratio to obtain the third-level mask area corresponding to the tower body rectangle.

[0071] In this embodiment, for each tower body rectangle in the transmission tower detection results, the variational mask overlap rate between this tower body rectangle and all tower base mask regions in the geographic environment segmentation results is calculated. The variational mask overlap rate is... The calculation formula is: In the formula, The mask image obtained after filling the rectangular frame of the tower body. For the tower base mask area, This represents the intersection operation. This represents a summation operation. The region filling algorithm can be an image flooding filling algorithm.

[0072] If among all the variable mask overlap rates corresponding to the rectangular frame of the tower body, there is a maximum variable mask overlap rate, and the maximum variable mask overlap rate is greater than the preset variable mask overlap rate threshold, then the tower base mask region corresponding to the maximum variable mask overlap rate is matched as the tower base mask region corresponding to the rectangular frame of the tower body.

[0073] If, among all the variable mask overlap rates corresponding to the tower body rectangle, there is no maximum variable mask overlap rate and the maximum variable mask overlap rate is greater than the preset variable mask overlap rate threshold, then the region of the tower body rectangle corresponding to the tower base rectangle in the transmission tower detection results is matched as the tower base mask region corresponding to the tower body rectangle.

[0074] Furthermore, for each rectangular frame of the transmission tower body in the transmission tower inspection results, in order to quantify the direct impact of the geographical environment of the transmission tower area on the transmission tower, the tower base mask area corresponding to the rectangular frame of the tower body is used as the first-level mask area corresponding to that rectangular frame of the tower body, which is used to focus on the geographical environment in direct contact with the tower foundation. Please refer to [link to relevant documentation]. Figure 3 The black polygon represents the geographic environment mask area, and the red polygon represents the first-level mask area.

[0075] Furthermore, to quantify the indirect impact of the surrounding geographical environment on the transmission tower, the tower base mask area corresponding to the rectangular frame of the tower body is expanded outward according to a first preset ratio to obtain a second-level mask area corresponding to the rectangular frame of the tower body. This second-level mask area is used to focus on the transitional environmental characteristics within a certain range around the tower. Please refer again. Figure 3 The green polygonal area is the second-level mask area.

[0076] In this embodiment, the first preset ratio can be 1:1. When expanding outward, the target size in the mask area remains unchanged, but the mask area becomes four times the area of ​​the tower base mask area. In other embodiments, the first preset ratio can also be other values, and this embodiment does not limit it.

[0077] Furthermore, to quantify the indirect impact of the surrounding geographical environment on the transmission towers, it is necessary to expand the tower base mask area corresponding to the rectangular frame of the tower body outward according to the second preset ratio, obtaining a third-level mask area corresponding to the rectangular frame of the tower body, which is used to reflect the overall environmental background of the large area where the tower is located. Please refer again. Figure 3 The blue polygonal area represents the third-level mask area.

[0078] The second preset ratio can be 2 times. When expanding outward, the target size in the mask area remains unchanged, but the mask area becomes 9 times the area of ​​the tower base mask area. In other embodiments, the second preset ratio can also be other values, and this embodiment does not limit it.

[0079] It should be noted that, based on the matched tower base mask area, the method expands outwards in stages according to a preset ratio to form multi-scale influence layers, simulating the near-to-medium-to-long-distance influence range of the geographical environment on transmission towers, demonstrating good physical interpretability. Furthermore, by using a preset expansion ratio instead of a fixed distance, the method adapts to towers of different sizes and their terrain variations, enhancing its universality and adaptability, and avoiding errors caused by a one-size-fits-all spatial division.

[0080] Step S104: Based on the degree of influence of environmental type, perform quantitative analysis on the first-level mask area, second-level mask area and third-level mask area corresponding to each of the rectangular frames of the tower body to obtain the first-level geographical environment influence factor, second-level geographical environment influence factor and third-level geographical environment influence factor corresponding to each of the rectangular frames of the tower body.

[0081] In this embodiment, since different geographical environment objects have different degrees of influence on transmission towers, the influence degree of environmental type is used to quantitatively analyze the first-level mask area, second-level mask area and third-level mask area corresponding to the rectangular frame of each tower body, and obtain multi-level geographical environment influence factors that characterize direct influence and indirect combined influence, including the first-level geographical environment influence factor, second-level geographical environment influence factor and third-level geographical environment influence factor corresponding to the rectangular frame of each tower body.

[0082] Step S105: Determine the degree of influence of the geographical environment on the transmission towers corresponding to the rectangular frames of each tower body based on the first-level geographical environment influence factor, the second-level geographical environment influence factor, and the third-level geographical environment influence factor.

[0083] In this embodiment, based on the numerical values ​​of the first-level, second-level, and third-level geographical environment influence factors of each tower's rectangular frame, the direct and indirect influence of the geographical environment on the transmission tower is comprehensively analyzed, thereby generating corresponding alarms, prompts, observations, and normal information, which can greatly improve the efficiency of power inspection work.

[0084] Please see Figure 4 In one specific embodiment, step S104 includes steps S1041 to S1045, and each step is described in detail below.

[0085] Step S1041: For each of the tower body rectangles, calculate the overlap rate of all geographic environment masks in the geographic environment segmentation result with the evolution masks of the first-level mask region, the second-level mask region, and the third-level mask region, respectively.

[0086] In this embodiment, to better quantify the influence between different geographic environment mask regions and different levels of mask regions for each transmission tower, the evolutionary mask overlap rate of all geographic environment masks in the geographic environment segmentation results with the first-level mask region, the second-level mask region, and the third-level mask region is calculated. The evolutionary mask overlap rate can represent the spatial interaction evolution characteristics during the dynamic expansion process, emphasizing the propagation and cumulative effects of environmental impacts, which aligns with the needs of long-term environmental evolution monitoring.

[0087] In one specific embodiment, step S1041 includes: for each of the geographic environment masks, calculating the intersection of the geographic environment mask with the first-level mask region, the second-level mask region, and the third-level mask region respectively; and determining the evolution mask overlap rate of the geographic environment mask with the first-level mask region, the second-level mask region, and the third-level mask region respectively based on the ratio of each intersection to the third-level mask region.

[0088] In this embodiment, for each geographic environment mask, the intersection of the geographic environment mask with the first-level mask region, the second-level mask region, and the third-level mask region is calculated; the ratio of each intersection to the third-level mask region is taken as the evolutionary mask overlap rate between the geographic environment mask and the first-level mask region, the second-level mask region, and the third-level mask region.

[0089] The specific formula for the evolution mask overlap rate is as follows: In the formula, For the first The environment type corresponding to each geographic environment mask For environment type The corresponding number Geographical mask With the The first rectangular frame of the tower body Level mask region The evolution of mask overlap rate.

[0090] It should be noted that, to ensure , , All three values ​​must be between 0 and 1; to ensure , , The three values ​​show an increasing trend; it is necessary to ensure the stability and rationality of the quantitative indicators.

[0091] In addition, due to The area contains area, The area contains The region, therefore Area As a fixed denominator, it ensures that the comparison quantity remains unchanged. As molecules, they characterize the geographic environment masking region and , , From the area of ​​their intersection, we can know that... Less than This ensures , , All are between 0 and 1.

[0092] At the same time, with the evolutionary calculations at different levels, , , The intersection area will also increase, thus ensuring , , The three values ​​show an increasing trend.

[0093] Meanwhile, as different levels of mask regions evolve outward, their quantitative indicators steadily increase, and these quantitative indicators also conform to the degree of impact in actual scenarios.

[0094] Step S1042: Based on the environment type corresponding to each of the geographical environment masks, preset multiple geographical environment degree influence coefficients; based on the different levels of mask areas corresponding to each of the tower body rectangles, preset multiple geographical environment basic influence coefficients.

[0095] In this embodiment, multiple geographical environment influence coefficients are preset according to environmental type, and multiple geographical environment basic influence coefficients are preset according to different mask areas, serving as multi-level standardized quantitative numerical indicators. This allows for the quantification of the impact on transmission towers by combining different environmental types and different geographical ranges.

[0096] As an example, assuming that the environment types include buildings, vegetation, water bodies, and floating objects, and the corresponding codes for each environment type are 2, 3, 4, and 5 respectively, then... ,in, For environment type The corresponding geographical environment influence coefficients show that the geographical environment influence coefficients for buildings, vegetation, water bodies, and floating objects are 1, 1, 2, and 3, respectively, while the geographical environment influence coefficients for other types are 0.

[0097] At the same time, assuming ,in, For the first The corresponding geographical environment basic influence coefficient.

[0098] Step S1043: Based on the multiple geographical environment influence coefficients and the evolution mask overlap rate, obtain the first-level geographical environment influence factor, the second-level geographical environment influence factor, and the third-level geographical environment influence factor corresponding to the rectangular frame of the tower body.

[0099] In this embodiment, by combining multiple geographical environment influence coefficients and evolution mask overlap rates, the first-level geographical environment influence factor, the second-level geographical environment influence factor, and the third-level geographical environment influence factor corresponding to the rectangular frame of the tower body are calculated to reflect the influence degree of different geographical environments.

[0100] The calculation formula is as follows: In the formula, The number of geographical environment masks, For environment type The corresponding geographical environment influence coefficient, For environment type The corresponding number Geographical mask With the The first rectangular frame of the tower body Level mask region The evolution of mask overlap rate.

[0101] Step S1044: Based on the multiple geographical environment basic influence coefficients and the evolution mask overlap rate, obtain the first-level geographical environment basic influence factor, the second-level geographical environment basic influence factor and the third-level geographical environment basic influence factor corresponding to the rectangular frame of the tower body.

[0102] In this embodiment, by combining multiple basic geographical environment influence coefficients and the evolution mask overlap rate of different levels of masks, the first-level, second-level, and third-level basic geographical environment influence factors corresponding to the rectangular frame of the tower body are calculated to reflect the influence attributes of different geographical environments.

[0103] The calculation formula is as follows: In the formula, For the first The corresponding geographical environment basic influence coefficient at level 1. This is the threshold for the overlap rate of the evolutionary mask. The calculation formula expresses that for any geographic environment mask, if the overlap rate between a geographic environment mask and the evolutionary masks of different levels of mask regions is greater than the evolutionary mask overlap rate threshold, the basic impact factors of different levels of geographic environment can be obtained. The corresponding basic impact factors of the first, second, and third levels of geographic environment are 10, 5, and 1, respectively. If no geographic environment mask has an overlap rate with the evolutionary masks of different levels of mask regions greater than the evolutionary mask overlap rate threshold, the basic impact factors of different levels of geographic environment cannot be obtained, and the corresponding basic impact factors of the first, second, and third levels of geographic environment are all 0.

[0104] Step S1045, according to the formula: The first-level, second-level, and third-level geographical environment impact factors corresponding to the rectangular frame of the tower body are calculated; where, For the first The first rectangular frame of the tower body Level 1 geographical environment influencing factors For the first The first rectangular frame of the tower body Influencing factors of geographical environment level No. The first rectangular frame of the tower body Level 1 Geographical environment basic influencing factors.

[0105] In this embodiment, the first-level geographical environment severity influence factor, the second-level geographical environment severity influence factor, and the third-level geographical environment severity influence factor are added to the first-level geographical environment basic influence factor, the second-level geographical environment basic influence factor, and the third-level geographical environment basic influence factor, respectively, to obtain the first-level geographical environment influence factor, the second-level geographical environment influence factor, and the third-level geographical environment influence factor.

[0106] The calculation formula is as follows: In the formula, For the first The first rectangular frame of the tower body Level 1 geographical environment influencing factors For the first The first rectangular frame of the tower body Influencing factors of geographical environment level For the first The first rectangular frame of the tower body Level 1 Geographical environment basic influencing factors.

[0107] In one specific embodiment, step S105 includes: for each of the pole body rectangles, if the first-level geographical environment impact factor corresponding to the pole body rectangle is greater than a first preset impact threshold, then the impact level is first level; if the first-level geographical environment impact factor corresponding to the pole body rectangle is 0, and the second-level geographical environment impact factor corresponding to the pole body rectangle is greater than the first preset impact threshold, then the impact level is second level; if the first-level geographical environment impact factor corresponding to the pole body rectangle is 0, and the second-level geographical environment impact factor corresponding to the pole body rectangle is greater than a second preset impact threshold and less than the first preset impact threshold, then the impact level is third level; if the first-level geographical environment impact factor corresponding to the pole body rectangle is 0, and the second-level geographical environment impact factor corresponding to the pole body rectangle is greater than a second preset impact threshold and less than the first preset impact threshold, then the impact level is third level; if the pole body rectangle corresponds to... If both the first-level and second-level geographical environment impact factors are 0, and the third-level geographical environment impact factor corresponding to the rectangular frame of the pole body is greater than the second preset impact threshold and less than the first preset impact threshold, then the impact level is level four. If both the first-level and second-level geographical environment impact factors corresponding to the rectangular frame of the pole body are 0, and the third-level geographical environment impact factor corresponding to the rectangular frame of the pole body is greater than the third preset impact threshold and less than the second preset impact threshold, then the impact level is level five. If all three levels of geographical environment impact factors corresponding to the rectangular frame of the pole body are 0, then the impact level is level six.

[0108] In this embodiment, the larger the first-level geographical environment influence factor, the greater the direct influence of the geographical environment on the tower base area of ​​the transmission tower; the larger the second-level geographical environment influence factor, the greater the indirect combined influence of the geographical environment on the tower base area of ​​the transmission tower; and the larger the third-level geographical environment influence factor, the greater the indirect combined influence of the geographical environment on the tower base area of ​​the transmission tower.

[0109] By setting multi-level impact determination rules, and performing combined logical judgments on the impact factors of each level of geographical environment based on multi-level impact thresholds, the direct impact of the geographical environment within the transmission tower area on the transmission tower, as well as the indirect impact of the geographical environment surrounding the transmission tower area on the transmission tower, are analyzed to meet the needs of differentiated handling strategies in power inspection.

[0110] Specifically, if the first-level geographical environment impact factor corresponding to the rectangular frame of the tower body is greater than the first preset impact threshold, it can be known that the tower base area of ​​the transmission tower has been directly affected by the geographical environment, and its corresponding impact level is the first level. An alarm message is generated for the transmission tower, and the first-level geographical environment impact factor is recorded.

[0111] If the first-level geographical environment impact factor corresponding to the rectangular frame of the tower body is 0, and the second-level geographical environment impact factor corresponding to the rectangular frame of the tower body is greater than the first preset impact threshold, then it can be known that the tower base area of ​​the transmission tower is not directly affected by the geographical environment, but is greatly affected by the indirect composite influence of the surrounding geographical environment. Its corresponding impact level is the second level. An alarm message is generated for the transmission tower, and the first-level geographical environment impact factor and the second-level geographical environment impact factor are recorded.

[0112] If the first-level geographical environment impact factor corresponding to the rectangular frame of the tower body is 0, and the second-level geographical environment impact factor corresponding to the rectangular frame of the tower body is greater than the second preset impact threshold and less than the first preset impact threshold, then it can be known that the tower base area of ​​the transmission tower is not directly affected by the geographical environment, but is greatly affected by the indirect composite impact of the surrounding geographical environment. Its corresponding impact level is the third level. A prompt message is generated for the transmission tower, and the first-level geographical environment impact factor and the second-level geographical environment impact factor are recorded.

[0113] If both the first-level and second-level geographical environment impact factors corresponding to the rectangular frame of the tower body are 0, and the third-level geographical environment impact factor corresponding to the rectangular frame of the tower body is greater than the second preset impact threshold but less than the first preset impact threshold, then it can be concluded that the tower base area of ​​the transmission tower is not directly affected by the geographical environment, but is significantly affected by the indirect combined influence of the surrounding geographical environment. Its corresponding impact level is fourth level. A prompt message is generated for the transmission tower, and the first-level, second-level, and third-level geographical environment impact factors are recorded. The impact level of the fourth level is less than that of the third level.

[0114] If the first-level and second-level geographical environment impact factors corresponding to the rectangular frame of the tower body are both 0, and the third-level geographical environment impact factor corresponding to the rectangular frame of the tower body is greater than the third preset impact threshold and less than the second preset impact threshold, then it can be known that the tower base area of ​​the transmission tower is not directly affected by the geographical environment, and is less affected by the indirect composite impact of the surrounding geographical environment. Its corresponding impact level is level 5. Observation information is generated for the transmission tower, and the first-level, second-level, and third-level geographical environment impact factors are recorded.

[0115] If the first-level, second-level, and third-level geographical environment impact factors corresponding to the rectangular frame of the tower body are all 0, it can be concluded that the tower base area of ​​the transmission tower is not directly affected by the geographical environment, nor is it indirectly affected by the surrounding geographical environment. Its corresponding impact level is level six. Normal information is generated for the transmission tower, and the first-level, second-level, and third-level geographical environment impact factors are recorded.

[0116] It should be noted that the first preset influence threshold can be 10, the second preset influence threshold can be 5, and the third preset influence threshold can be 1. In other embodiments, they can also be other values, which are not limited in this embodiment.

[0117] The remote sensing image analysis method for the geographical environment of power transmission towers proposed in this embodiment constructs a fully automated remote sensing image analysis process, including remote sensing detection of power transmission towers, geographical environment mask segmentation, mask region matching and expansion, quantitative analysis of the geographical environment, and determination of the degree of influence. It overcomes the problems of low detection accuracy and difficulty in quantifying environmental impact caused by small targets and complex backgrounds in high-resolution remote sensing images, which are common with traditional machine vision algorithms and general deep learning models. It realizes the quantitative analysis of the degree of influence of the surrounding geographical environment on power transmission towers, and greatly improves the efficiency of power inspection work.

[0118] Example 2

[0119] Furthermore, this embodiment proposes a remote sensing image analysis system 500 for the geographical environment of power transmission towers. Please refer to [link to relevant documentation]. Figure 5 The system includes:

[0120] The target detection module 501 is used to acquire remote sensing images and input the remote sensing images into the transmission tower remote sensing detection algorithm to obtain the transmission tower detection results.

[0121] The mask segmentation module 502 is used to input the remote sensing image into the geographic environment remote sensing segmentation algorithm to obtain the geographic environment segmentation result if there is a rectangular frame of the tower body in the tower detection result.

[0122] The matching expansion module 503 is used to perform mask region matching and mask region expansion based on the transmission tower detection results and the geographical environment segmentation results if there is a geographical environment mask region in the geographical environment segmentation results, so as to obtain the first-level mask region, the second-level mask region and the third-level mask region corresponding to each of the tower body rectangles;

[0123] The quantitative analysis module 504 is used to perform quantitative analysis on the first-level mask area, the second-level mask area and the third-level mask area corresponding to the rectangular frame of each tower body based on the degree of influence of environmental type, so as to obtain the first-level geographical environment influence factor, the second-level geographical environment influence factor and the third-level geographical environment influence factor corresponding to the rectangular frame of each tower body.

[0124] The degree determination module 505 is used to determine the degree of influence of the geographical environment on the transmission towers corresponding to the rectangular frames of each tower body based on the first-level geographical environment influence factor, the second-level geographical environment influence factor and the third-level geographical environment influence factor corresponding to each of the tower body rectangular frames.

[0125] The remote sensing image analysis system for the geographical environment of power transmission towers proposed in this embodiment constructs a fully automated remote sensing image analysis process, including remote sensing detection of power transmission towers, geographical environment mask segmentation, mask region matching and expansion, quantitative analysis of the geographical environment, and determination of the degree of influence. It overcomes the problems of low detection accuracy and difficulty in quantifying environmental impact caused by small targets and complex backgrounds in high-resolution remote sensing images, which are common with traditional machine vision algorithms and general deep learning models. It realizes the quantitative analysis of the degree of influence of the surrounding geographical environment on power transmission towers, and greatly improves the efficiency of power inspection work.

[0126] Example 3

[0127] This disclosure also provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the remote sensing image analysis method for the geographical environment of transmission towers described in Embodiment 1.

[0128] Example 4

[0129] This disclosure also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the remote sensing image analysis method for the geographical environment of transmission towers described in Embodiment 1.

[0130] In all examples shown and described herein, any specific values ​​should be interpreted as merely exemplary and not as limitations; therefore, other examples of exemplary embodiments may have different values.

[0131] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0132] The above embodiments are merely illustrative of several implementations of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the protection scope of the present invention.

Claims

1. A method for analyzing remote sensing images of the geographical environment of power transmission towers, characterized in that, include: Acquire remote sensing images and input them into the remote sensing detection algorithm for power transmission towers to obtain the detection results of the power transmission towers; If the tower body rectangle exists in the tower detection result, the remote sensing image is input into the geographic environment remote sensing segmentation algorithm to obtain the geographic environment segmentation result. If there is a geographic environment mask region in the geographic environment segmentation result, then the mask region matching and mask region expansion are performed according to the transmission tower detection result and the geographic environment segmentation result to obtain the first-level mask region, the second-level mask region and the third-level mask region corresponding to the rectangular frame of each tower body; Based on the degree of influence of environmental type, a quantitative analysis is performed on the first-level mask area, the second-level mask area and the third-level mask area corresponding to the rectangular frame of each tower body to obtain the first-level geographical environment influence factor, the second-level geographical environment influence factor and the third-level geographical environment influence factor corresponding to the rectangular frame of each tower body. The degree of influence of the geographical environment on the transmission towers corresponding to the rectangular frames of each tower body is determined based on the first-level geographical environment influence factor, the second-level geographical environment influence factor and the third-level geographical environment influence factor corresponding to the rectangular frames of each tower body. Based on the degree of influence of environmental type, a quantitative analysis is performed on the first-level mask area, second-level mask area, and third-level mask area corresponding to the rectangular frame of each tower body to obtain the first-level geographical environment influence factor, second-level geographical environment influence factor, and third-level geographical environment influence factor corresponding to the rectangular frame of each tower body, including: For each of the aforementioned tower body rectangles, calculate the overlap rate between all geographic environment masks in the geographic environment segmentation results and the evolution masks of the first-level mask region, the second-level mask region, and the third-level mask region, respectively; Based on the environment type corresponding to each of the aforementioned geographical environment masks, multiple geographical environment degree influence coefficients are preset; based on the different levels of mask areas corresponding to each of the aforementioned tower body rectangles, multiple geographical environment basic influence coefficients are preset. Based on the multiple geographical environment influence coefficients and the evolution mask overlap rate, the first-level geographical environment influence factor, the second-level geographical environment influence factor, and the third-level geographical environment influence factor corresponding to the rectangular frame of the tower body are obtained. Based on the multiple geographical environment basic influence coefficients and the evolution mask overlap rate, the first-level geographical environment basic influence factor, the second-level geographical environment basic influence factor and the third-level geographical environment basic influence factor corresponding to the rectangular frame of the tower body are obtained. According to the formula: The first-level geographical environment impact factor, the second-level geographical environment impact factor, and the third-level geographical environment impact factor corresponding to the rectangular frame of the tower body are calculated. In the formula, For the first The first rectangular frame of the tower body Level 1 geographical environment influencing factors For the first The first rectangular frame of the tower body Influencing factors of geographical environment level No. The first rectangular frame of the tower body Level 1 Geographical environment basic influencing factors; The calculation formulas for the first-level, second-level, and third-level geographical environment basic impact factors corresponding to the rectangular frame of the tower body are as follows: In the formula, For the first The corresponding geographical environment basic influence coefficient at level 1. The number of the geographic environment masks. For environment type The corresponding number Geographical mask With the The first rectangular frame of the tower body Level mask region The evolution of mask overlap rate, This is the threshold for the overlap rate of the evolution mask; Specifically, for each of the rectangular frames of the tower body, calculating the overlap rate of all geographic environment masks in the geographic environment segmentation result with the evolution masks of the first-level mask region, the second-level mask region, and the third-level mask region includes: For each of the aforementioned geographic environment masks, the intersections between the geographic environment mask and the first-level mask region, the second-level mask region, and the third-level mask region are calculated respectively; Based on the ratio of each intersection to the third-level mask region, the overlap rate of the geographic environment mask with the evolution masks of the first-level mask region, the second-level mask region, and the third-level mask region is determined. The step of matching and expanding the mask region based on the transmission tower detection results and the geographical environment segmentation results to obtain the first-level mask region, the second-level mask region, and the third-level mask region corresponding to the rectangular frame of each tower body includes: The detection results of the power transmission towers and the geographical environment segmentation results are matched to obtain the tower base mask area corresponding to the rectangular frame of each tower body; For each of the aforementioned tower body rectangular frames, the tower base mask area corresponding to the tower body rectangular frame is taken as the first-level mask area corresponding to the tower body rectangular frame; The tower base mask area is expanded outward according to the first preset ratio to obtain the second-level mask area corresponding to the rectangular frame of the tower body. The tower base mask area is expanded outward according to the second preset ratio to obtain the third-level mask area corresponding to the rectangular frame of the tower body.

2. The method for analyzing remote sensing images of the geographical environment of transmission towers according to claim 1, characterized in that, The calculation formulas for the first-level, second-level, and third-level geographical environment influence factors corresponding to the rectangular frame of the tower body are as follows: In the formula, For environment type The corresponding geographical environment influence coefficient.

3. The method for analyzing remote sensing images of the geographical environment of transmission towers according to claim 1, characterized in that, The formula for calculating the overlap rate of the evolution mask is as follows: In the formula, This represents the summation operation. This represents the intersection operation.

4. The method for analyzing remote sensing images of the geographical environment of transmission towers according to claim 1, characterized in that, The determination of the degree of influence of the geographical environment on the transmission towers corresponding to the rectangular frames of each tower body based on the first-level, second-level, and third-level geographical environment influence factors includes: For each of the aforementioned pole body rectangles, if the first-level geographical environment impact factor corresponding to the pole body rectangle is greater than the first preset impact threshold, then the degree of impact is the first level. If the first-level geographical environment impact factor corresponding to the rectangular frame of the pole body is 0, and the second-level geographical environment impact factor corresponding to the rectangular frame of the pole body is greater than the first preset impact threshold, then the degree of impact is the second level. If the first-level geographical environment impact factor corresponding to the rectangular frame of the tower body is 0, and the second-level geographical environment impact factor corresponding to the rectangular frame of the tower body is greater than the second preset impact threshold and less than the first preset impact threshold, then the degree of impact is the third level. If the first-level geographical environment impact factor and the second-level geographical environment impact factor corresponding to the rectangular frame of the pole body are both 0, and the third-level geographical environment impact factor corresponding to the rectangular frame of the pole body is greater than the second preset impact threshold and less than the first preset impact threshold, then the degree of impact is the fourth level. If the first-level geographical environment impact factor and the second-level geographical environment impact factor corresponding to the rectangular frame of the pole body are both 0, and the third-level geographical environment impact factor corresponding to the rectangular frame of the pole body is greater than the third preset impact threshold and less than the second preset impact threshold, then the degree of impact is the fifth level. If the first-level, second-level, and third-level geographical environment impact factors corresponding to the rectangular frame of the tower body are all 0, then the degree of impact is level six.

5. The method for analyzing remote sensing images of the geographical environment of transmission towers according to claim 1, characterized in that, The remote sensing detection algorithm for power transmission towers is an improved YOLO11 detection algorithm. This improved YOLO11 detection algorithm includes a backbone network, a feature pyramid network, and a head detection network. The feature pyramid network includes a first self-attention convolutional module, a second self-attention convolutional module, a first convolutional layer, a second convolutional layer, a first SwinTransformer module, a second SwinTransformer module, and a third SwinTransformer module. The head detection network includes a first classification regression module, a second classification regression module, a third classification regression module, a fourth classification regression module, a fifth classification regression module, a sixth classification regression module, a first interest alignment module, a second interest alignment module, and a third interest alignment module. The process of inputting the remote sensing image into the power transmission tower remote sensing detection algorithm to obtain the power transmission tower detection results includes: The remote sensing image is used to extract features through the backbone network to obtain a first layer feature map, a second layer feature map and a third layer feature map; The third-layer feature map and the second-layer feature map are input into the first self-attention convolution module to obtain the first self-attention convolution feature map, and the first self-attention convolution feature map and the second-layer feature map are concatenated to obtain the first fusion map. The first fused image and the first layer feature map are input into the second self-attention convolution module to obtain the second self-attention convolution feature map, and the second self-attention convolution feature map is concatenated with the first layer feature map to obtain the second fused image; The second fusion map is input into the first convolutional layer to obtain the first feature map, and the first feature map is concatenated with the first fusion map to obtain the third fusion map; The third fusion map is input into the second convolutional layer to obtain the second feature map, and the second feature map is concatenated with the third layer feature map to obtain the fourth fusion map; The second fused image is input into the first SwinTransformer module to obtain a first fine-grained feature map; the third fused image is input into the second SwinTransformer module to obtain a second fine-grained feature map; and the fourth fused image is input into the third SwinTransformer module to obtain a third fine-grained feature map. The first fine-grained feature map, the second fine-grained feature map, and the third fine-grained feature map are respectively input into the first classification and regression module, the second classification and regression module, and the third classification and regression module to obtain the first tower body detection result, the second tower body detection result, and the third tower body detection result; The first tower body detection result, the second tower body detection result, and the third tower body detection result are respectively input into the first interest alignment module, the second interest alignment module, and the third interest alignment module to obtain the first alignment feature map, the second alignment feature map, and the third alignment feature map; The first alignment feature map, the second alignment feature map, and the third alignment feature map are respectively input into the fourth classification regression module, the fifth classification regression module, and the sixth classification regression module to obtain the first tower base detection result, the second tower base detection result, and the third tower base detection result; The transmission tower inspection results are obtained based on the inspection results of the first tower body, the second tower body, the third tower body, the first tower base, the second tower base, and the third tower base.

6. A remote sensing image analysis system for the geographical environment of power transmission towers, characterized in that, include: The target detection module is used to acquire remote sensing images and input the remote sensing images into the transmission tower remote sensing detection algorithm to obtain the transmission tower detection results. The mask segmentation module is used to input the remote sensing image into the geographic environment remote sensing segmentation algorithm to obtain the geographic environment segmentation result if the tower body rectangle exists in the tower detection result. The matching expansion module is used to match and expand the mask area according to the transmission tower detection result and the geographical environment segmentation result if there is a geographical environment mask area in the geographical environment segmentation result, so as to obtain the first-level mask area, the second-level mask area and the third-level mask area corresponding to the rectangular frame of each tower body; The quantitative analysis module is used to perform quantitative analysis on the first-level mask area, second-level mask area and third-level mask area corresponding to the rectangular frame of each tower body based on the degree of influence of environmental type, so as to obtain the first-level geographical environment influence factor, second-level geographical environment influence factor and third-level geographical environment influence factor corresponding to the rectangular frame of each tower body. The degree determination module is used to determine the degree of influence of the geographical environment on the transmission towers corresponding to the rectangular frames of each tower body based on the first-level geographical environment influence factor, the second-level geographical environment influence factor and the third-level geographical environment influence factor corresponding to the rectangular frames of each tower body. Based on the degree of influence of environmental type, a quantitative analysis is performed on the first-level mask area, second-level mask area, and third-level mask area corresponding to the rectangular frame of each tower body to obtain the first-level geographical environment influence factor, second-level geographical environment influence factor, and third-level geographical environment influence factor corresponding to the rectangular frame of each tower body, including: For each of the aforementioned tower body rectangles, calculate the overlap rate between all geographic environment masks in the geographic environment segmentation results and the evolution masks of the first-level mask region, the second-level mask region, and the third-level mask region, respectively; Based on the environment type corresponding to each of the aforementioned geographical environment masks, multiple geographical environment degree influence coefficients are preset; based on the different levels of mask areas corresponding to each of the aforementioned tower body rectangles, multiple geographical environment basic influence coefficients are preset. Based on the multiple geographical environment influence coefficients and the evolution mask overlap rate, the first-level geographical environment influence factor, the second-level geographical environment influence factor, and the third-level geographical environment influence factor corresponding to the rectangular frame of the tower body are obtained. Based on the multiple geographical environment basic influence coefficients and the evolution mask overlap rate, the first-level geographical environment basic influence factor, the second-level geographical environment basic influence factor and the third-level geographical environment basic influence factor corresponding to the rectangular frame of the tower body are obtained. According to the formula: The first-level geographical environment impact factor, the second-level geographical environment impact factor, and the third-level geographical environment impact factor corresponding to the rectangular frame of the tower body are calculated. In the formula, For the first The first rectangular frame of the tower body Level 1 geographical environment influencing factors For the first The first rectangular frame of the tower body Influencing factors of geographical environment level No. The first rectangular frame of the tower body Level 1 Geographical environment basic influencing factors; The calculation formulas for the first-level, second-level, and third-level geographical environment basic impact factors corresponding to the rectangular frame of the tower body are as follows: In the formula, For the first The corresponding geographical environment basic influence coefficient at level 1. The number of the geographic environment masks. For environment type The corresponding number Geographical mask With the The first rectangular frame of the tower body Level mask region The evolution of mask overlap rate, This is the threshold for the overlap rate of the evolution mask; Specifically, for each of the rectangular frames of the tower body, calculating the overlap rate of all geographic environment masks in the geographic environment segmentation result with the evolution masks of the first-level mask region, the second-level mask region, and the third-level mask region includes: For each of the aforementioned geographic environment masks, the intersections between the geographic environment mask and the first-level mask region, the second-level mask region, and the third-level mask region are calculated respectively; Based on the ratio of each intersection to the third-level mask region, the overlap rate of the geographic environment mask with the evolution masks of the first-level mask region, the second-level mask region, and the third-level mask region is determined. The step of matching and expanding the mask region based on the transmission tower detection results and the geographical environment segmentation results to obtain the first-level mask region, the second-level mask region, and the third-level mask region corresponding to the rectangular frame of each tower body includes: The detection results of the power transmission towers and the geographical environment segmentation results are matched to obtain the tower base mask area corresponding to the rectangular frame of each tower body; For each of the aforementioned tower body rectangular frames, the tower base mask area corresponding to the tower body rectangular frame is taken as the first-level mask area corresponding to the tower body rectangular frame; The tower base mask area is expanded outward according to the first preset ratio to obtain the second-level mask area corresponding to the rectangular frame of the tower body. The tower base mask area is expanded outward according to the second preset ratio to obtain the third-level mask area corresponding to the rectangular frame of the tower body.

7. A computer device, characterized in that, The method includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the remote sensing image analysis method for the geographic environment of transmission towers as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the remote sensing image analysis method for the geographic environment of transmission towers as described in any one of claims 1-5.

Images